Molecular imaging is an emerging technology that can provide information beyond physical appearance of living tissues by detecting molecular signatures in a noninvasive way. This technology can be highly useful for diagnosis, biopsy or surgical guidance, and optimization of preclinical and clinical tests of medication. Among various molecular imaging technologies, Raman spectroscopy is particularly interesting because it is safe, inexpensive, agent-free, and most importantly, capable of detecting multiple molecules simultaneously. Ordinary Raman spectroscopy, however, has a significant drawback, i.e. low sensitivity. Placing nanometallic structures in close proximity with molecules can augment the Raman signal enormously. Utilization of this surface-enhanced Raman spectroscopy (SERS) has a great potential for the molecular imaging. However, it has been only demonstrated in a setting of an external Raman microscope with nanoparticles injected into small animals. Although these demonstrations show promises of in situ SERS, to implement the method in a clinical setting, it has to be used in conjunction with endoscopy to circumvent the penetration limit of light. As a beginning effort along that line, the project will examine the feasibility of in situ contact-type endoscopic SERS. The project will begin with fabrication of nanometallic structures for SERS. More specifically, four different methods will be attempted, aiming at large surface enhancement, good uniformity, and low fabrication cost. Those methods are 1) nano-molding with nanoporous silicon, 2) nano-molding with nanospheres, 3) mask shape transfer etching after nanosphere lithography, and 4) room-temperature deposition of carbon nanotube (CNT) film. The SERS substrates will be characterized in various aspects, and compared to one another. The endoscopic SERS probe head will be developed, which is comprised of a SERS substrate, a graded- index lens, an optical fiber, and a housing. These components will be assembled and packaged after alignment. Its characteristics will be examined by using a Raman spectrometer system built in-house, which consists of a laser, a dichroic mirror, lenses, a linear CCD array, and a computer. Developed SERS probes will become a core component for the next phase of research, starting from ex vivo SERS imaging. Since the enhancement of SERS shows strong dependence on the distance between the target molecules and the nanometallic surfaces, it is critically important to examine the effect of interfering materials to the imaging, to evaluate the feasibility of the proposed method. For this, Raman study of the Barrett's esophagus by Shetty et al will be used as a model, which identified DNA, glycogen, oleic acid, and actin as diagnostic signature substances for the disease. Viscous fluids and/or self-assembled monolayers will be placed between the target molecules and SERS probes, simulating interfering materials. Minimum detectable concentration level will be studied for each type of molecules. A mixture of molecules with various mixing ratio will be examined also with a SERS probe and the principal component analysis to assess the accuracy of the method.